A signal x(t) is bandlimited to B Hz

  • Thread starter Thread starter gabel
  • Start date Start date
  • Tags Tags
    Signal
AI Thread Summary
A signal x(t) that is bandlimited to B Hz can be shown to have its n-th power, x^n(t), bandlimited to nB Hz. The Fourier transform of a product of two functions is their convolution, which maintains the bandlimit properties of the original signals. By taking two bandlimited signals A(f) and B(f) with limits B1 and B2, the convolution A * B(f) results in a new bandlimited signal. The conditions for the integrand being non-zero restrict the values of f, confirming that the convolution remains bandlimited. Mathematical induction can be applied to demonstrate that the n-th power of a bandlimited signal retains this property, ultimately proving that x^n(t) is bandlimited to nB Hz.
gabel
Messages
17
Reaction score
0
A singal x(t) is bandlimited to B Hz. Show that that the signal x^{n}(t) is bandlimited to nB Hz.

I have no idee on how to adress this problem. Can get some help?
 
Physics news on Phys.org


A band-limited signal has a Fourier component containing a heaviside step function:

<br /> X(f) = c(f) \theta(B - |f|)<br />

The Fourier transform of a product of two functions is their convolution:

<br /> \begin{array}{l}<br /> \mathrm{F.T.}[a(t) b(t)](f) = \int_{-\infty}^{\infty}{a(t) \, b(t) \, e^{-2\pi j f t} \, dt} \\ \\<br /> <br /> = \int_{-\infty}^{\infty}{dt \, e^{-2\pi j f t} \, \int_{-\infty}^{\infty}{A(f_{1}) \, e^{2\pi j f_{1} t} \, df_{1}} \, \int_{-\infty}^{\infty}{B(f_{2}) \, e^{2\pi j f_{2} t} \, df_{2}}} \\ \\<br /> <br /> = \int_{-\infty}^{\infty}{\int_{-\infty}^{\infty}{df_{1} \, df_{2} \, A(f_{1}) \, B(f_{2}) \, \int_{-\infty}^{\infty}{dt \, e^{2\pi j (f_{1} + f_{2} - f) t} \, dt}}} \\ \\<br /> <br /> = \int_{-\infty}^{\infty}{\int_{-\infty}^{\infty}{df_{1} \, df_{2} \, A(f_{1}) \, B(f_{2}) \, \delta(f_{1} + f_{2} - f)} \\ \\<br /> <br /> = \int_{-\infty}^{\infty}{A(f&#039;) \, B(f - f&#039;) \, df&#039;}<br /> \end{array}<br />

Take the signals A(f) and B(f) to be band-limited with band limits B_{1} and B_{2}, respectively. Then, their convolution is:

<br /> A \ast B(f) = \int_{-\infty}^{\infty}{A(f&#039;) \, \theta(B_{1} - |f&#039;|) \, B(f - f&#039;) \, \theta(B_{2} - |f - f&#039;|) \, df&#039;}<br />

The integrand is non-zero only when:

<br /> \left\{\begin{array}{lcl}<br /> B_{1} - |f&#039;| &amp; \ge &amp; 0 \\<br /> <br /> B_{2} - |f - f&#039;| &amp; \ge &amp; 0<br /> \end{array}\right.<br />

These conditions limit the domain of integration with respect to f&#039;. But, they also give some necessary conditions on the possible values of f when the above conditions are not contradictory. This makes the convolution also band limited. What is the band limit on the convolution in terms of B_{1} and B_{2}?

Then, use Mathematical Induction to prove the band limit of the n-th power.
 

Similar threads

Replies
4
Views
2K
Replies
15
Views
59K
Replies
1
Views
2K
Replies
3
Views
2K
Replies
1
Views
2K
Replies
6
Views
5K
Replies
4
Views
1K
Back
Top